论文部分内容阅读
为实现无人作战飞机(UCAV)认知导航的高鲁棒性特征点提取,提出一种基于自适应非极大值抑制(ANMS)的多元量化Hessian-Affine迭代式尺度不变特征变换(SIFT)方法。针对认知导航对特征点均匀分布的需求,提出基于ANMS的初始特征点优选算法。为确保特征点的仿射不变性,利用引入迭代调节因子的Hessian-Affine迭代算法估计仿射不变区域,并在对应归一化圆形区域进行主方向确定以及圆形描述子生成。针对模拟特征序列分布不均匀、正确匹配率不高的缺陷,采用多值量化与比特抽取结合法对模拟特征序列进行多元量化,并且分析验证了该方法的优越性能。仿真结果表明,本文方法具有较高的正确匹配率,具有旋转不变性和尺度不变性,其抗噪性能提高了10dB,并且在大视角变化范围内具有较优的抗仿射性能。
In order to extract highly robust feature points of UCAV cognitive navigation, a multivariate quantized Hessian-Affine iterative scaling invariant feature transform based on adaptive non-maximal-valued suppression (ANMS) is proposed (SIFT )method. In order to meet the demand of uniform distribution of feature points in cognitive navigation, an initial feature point optimization algorithm based on ANMS is proposed. In order to ensure the affine invariance of the feature points, the Hessian-Affine iterative algorithm which introduces the iterative adjustment factor is used to estimate the affine invariant regions and the main directions are determined and the circle descriptor is generated in the corresponding normalized circular region. Aiming at the defect that the simulation feature sequence is unevenly distributed and the correct matching rate is not high, the multivalued quantization and bit extraction combination method are used to multivariate the simulation feature sequences. The superiority of this method is verified by analysis. The simulation results show that the proposed method has high correct matching rate, rotation invariance and scale invariance, and its anti-noise performance is improved by 10dB. It also has better anti-affine performance in the range of large viewing angle.